Re: new test cases (long)

[Mark Wielaard <mark at klomp dot org>]

Hi Raif,

Thanks for all the pointers. The character encoding names seem to be
confusing whatever way you look at it. What is and isn't a canonical
name, for what package, what the (historical) alias is, etc is difficult
to decipher.

Also note that the 1.4 docs and 1.4.1 encoding docs actually list
different canonical names... Duh...

Anyway I think the best thing todo is to add all canonical, historical
and/or alias must support character names to the getBytes() tests, at
least for the names that are documented on all these different
(versions) of the API/Spec pages. The distinction between names used for
java.lang/io and java.nio seems to only confuse matters and
implementations that only support some names for some of the library
classes will probably confuse users enormously.

So getBytes14 now tests US-ASCII, windows-1252, ISO-8859-1, ISO-8859-15,
ISO8859_15, UTF-16BE and UTF-16LE. Together with the getBytes13 tests
this should catch all the encoding names that people will probaly always
expect to be available in a normal class library implementation.

Cheers,

Mark